SciELO - Scientific Electronic Library Online

vol.30 issue3Comparative analysis of the trypanocidal activity and chemical properties of E-lychnophoric acid and its derivatives using theoretical calculationsElectrochemical evidence of sulfinic acid derivative as an intermediate in the reduction of aromatic sulfonyl chloride in an aprotic medium author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand



  • English (pdf)
  • English (epdf)
  • Article in xml format
  • Article references
  • How to cite this article
  • SciELO Analytics
  • Curriculum ScienTI
  • Automatic translation


Related links


Eclética Química

Print version ISSN 0100-4670On-line version ISSN 1678-4618

Eclet. Quím. vol.30 no.3 São Paulo July/Sept. 2005 

Determination of Hg in water by CVAAS using 2-aminothiazole modified silica


Determinação de mercúrio em águas por CVAAS empregando sílica modificada por 2-aminotiazol



F. A. SilvaI; I. L. de AlcântaraII; P. S. RoldanII; C. C. F. PadilhaIII; A. B. de AraújoII; J. P. S. ValenteI; A. O. FlorentinoI; P. M. PadilhaI, *

IInstituto de Biociências – Departamento de Química e Bioquímica/UNESP Caixa Postal 510, 18618-000, Botucatu, SP, Brazil
IIInstituto de Química – Departamento de Química Analítica/UNESP Caixa Postal 355, 14800-900, Araraquara, SP, Brazil
IIIInstituto de Biociências – Departamento de Física e Biofísica/UNESP Caixa Postal 510, 18618-000, Botucatu, SP, Brazil




This paper discusses a rapid and sensitive method developed to determine trace levels of mercury in natural water samples by cold vapor atomic absorption spectrometry using a preconcentration system composed by mini-column packed with 100 mg of 2-aminothiazol modified silica gel (SiAT) coupled on-line with the spectrometer's cold vapor generator system. The optimum preconcentration conditions are also described here. The preconcentrated Hg(II) ions were eluted directly from the column to the spectrometer's cold vapor generator system using 100 µL of 2 mol L-1 hydrochloric acid and the retention efficiency achieved exceeded 95%. The enrichment factors determined were 29, 38 and 46 using 3, 4 and 5 mL of preconcentrated aqueous solutions containing 400 ng L-1 of Hg. The detection limit calculated was 5 ng L-1. The preconcentration procedure was applied to determine trace level mercury in spiked river water samples.

Keywords: mercury; flow-injection analysis; CVAAS; modified silica; preconcentration.


O presente trabalho descreve o desenvolvimento de um método rápido e sensível para determinar traços de mercúrio em amostras de águas naturais por espectrometria de absorção atômica com geração de vapor a frio, utilizando-se um sistema de preconcentração composto por uma mini-coluna empacotada com 100 mg de sílica gel modificada com o ligante 2-aminotiazol e acoplado em linha com o sistema de geração de vapor a frio do espectrômetro de absorção atômica. As condições ótimas de preconcentração são descritas. Os íons Hg(II) preconcentrados foram eluídos diretamente da mini-coluna para o sistema de geração de vapor a frio utilizando-se 100 µL de solução 2 mol L-1 de ácido clorídrico, com uma eficiência de 95%. Os fatores de enriquecimento determinados foram 29, 38 e 46 percolando-se 3, 4 e 5 mL de solução aquosa contendo 400 ng L-1 de Hg(II). O limite de detecção calculado foi de 5 ng L-1. O procedimento de preconcentração foi aplicado na determinação de traços de mercúrio adicionados em amostras de águas de rio.

Palavras-chave: mercúrio; análise por injeção em fluxo; CVAAS; sílica modificada; preconcentração.




Heavy metals (e.g., Hg) at trace levels are provenly toxic elements and can cause human health disorders [1-3]. Since the main sources of mercury for humans are water and food, the monitoring of mercury content in natural waters is of paramount importance [4-6].

Mercury contamination in environmental samples is usually caused by less than 1 mg L-1 of this element, thus requiring highly sensitive techniques to determine and quantify its presence. Several analytical techniques exist to determine mercury at trace levels [7-20]. One of these techniques is cold vapor atomic absorption spectrometry (CVAAS), which is suitable due to its high sensitivity for accurate determinations of mercury [11, 14, 18]. However, CVAAS technique's selectivity and sensitivity to environmental samples is strongly dependent on the matrix components [19]. Preconcentration steps based on amalgamation onto a gold trap have been used to enhance the sensitivity in mercury determinations by atomic absorption spectrometry [7-10, 15, 16]. The flow injection preconcentration technique is useful for application in determining mercury by CVAAS, providing improved performance by enhancing the sensitivity, separating the analyte from the matrix, and decreasing of the risk of contamination [13, 18, 20].

In recent years, studies have focused on the use of chemically modified silica gel matrices with various groups to chelate organofunctional groups designed for the adsorption and preconcentration of metal ions in aqueous media [21-27]. The literature contains descriptions of columns packed with chemically modified silica gel with groups attached having a selective affinity for mercury ions [22-26]. A manual procedure with this sorbent packed column has been proposed for pre-concentrating mercury [25, 26]. However, the routine applicability of this procedure is limited by the time-consuming analysis it involves (100 min for each determination).

This paper describes the analytical performance of a flow injection system with on-line preconcentration in a sorbent mini-column packed with 2-aminothiazole modified silica (SiAT) gel for determining mercury in a series of spiked river water samples. The system proposed here was tested initially on a synthetic aqueous solution containing trace levels of mercury, and was subsequently applied to spiked river water samples.


Materials and methods

Reagents, analytical solutions and samples

All the solutions were prepared with high purity chemicals and distilled-deionized water (Milli-Q system, Millipore). The hydrochloric and nitric acids used were Suprapur (Merck).

A solution of 10% m/v SnCl2 in 6 mol L-1 HCl was prepared daily by dissolving the appropriate amount of the salt (Merck) in about 50 mL of concentrated hydrochloric acid and diluting it in up to 100 mL of water.

Standard mercury solutions (80 – 800 ng L-1) were prepared by stepwise dilution from 1000 mg L-1 stock solution (Merck – Titrisol) with Milli-Q water, acidified with nitric acid to pH 2.

The 2-aminothiazole modified silica (SiAT) was prepared as described by Alcântara et al. [28]. The column (a 5 cm long glass tube with 3 mm i.d.) was dry-packed with 100 mg of SiAT.

Samples collected from natural water (Tietê, Parapanema and Paraná rivers, SP, Brazil) were immediately filtered through a 0.45 µm membrane. After acidification (pH < 2.0) with 6.0 mol L-1 HNO3 solution, the samples were stored in highly purified PE containers. A volume of 50 mL water sample and 30 µL of 1000 µg L-1 Hg(II) were added to a 100 mL volumetric flask and the resulting solution diluted to the final volume with water. The river water samples spiked with 300 ng L-1 Hg(II) were digested by photocatalysis in the presence of 100 mg of TiO2 and 0.01 mol L-1 potassium persulfate.

Mineralization of river water samples spiked with Hg(II)

The mineralization of river water samples spiked with 300 ng L-1 Hg(II) were performed in a Quimis cylindrical photoreactor thermostatted at 25°C (298K) with luminous intensity of 3.81mW/cm2 . Nitrogen was bubbled into the sample before conducting the experiment to remove excess oxygen. A BECKMAN model PMA 2100 radiometer was used to measure the luminous intensity. A mass of 100 mg of TiO2 and 2,70 g K2S2O8 were suspended in 100 mL of samples and stirred for 60 min under mechanical agitation. The photocatalytic reaction was conducted with the reactor open to the air. After the period of time, the TiO2 was separated of solution by fitration using 0.22µm Millipore filters and the organic carbon concentration was determined using Total Organic Carbon Analizer SHIMADZU/TOC-VCPH.


A SHIMADZU model AA-6800 atomic absorption spectrometer equipped with a hollow mercury cathode lamp and a deuterium lamp for background correction, and a homemade T-shaped acrylic flow cell with quartz windows (190 x 4 mm) were used. The spectrometer's monochromator was adjusted to 253.7 nm (the resonance line corresponding to a wavelength of great sensitivity for mercury) and the spectral-resolution to 0.5 nm. The atomic signal was measured in peak height mode.

An Ismatec model IPC-8 peristaltic pump equipped with Tygon pumping tubes (i.d. 0.5 mm) and a homemade Perspex injector-commutator were used in the flow injection system.

Preconcentration system

Figure 1 shows the flow diagram of the SiAT-packed mini-column system, which consisted of a peristaltic pump (1), an injection valve-1 (2), a mini-column packed with SiAT (4), a Loop-LR (5), a reaction cell (6) coupled to the AAS flow cell (10), and a N2 line (9) crossing the injection valve 3 (8). Standard solutions or water samples in the 80 – 800 ng L-1 Hg(II) concentration range were passed through the column at 0.5 mL min-1 for 4 min. to sorb the Hg(II) in the column with the opening valve-1 (2). The mercury was eluted with 100 µL of 2.0 mol L-1 HCl solution and added through the injection valve-1 (2) at 1 mL min-1. The eluted Hg(II) ions were transported to the reaction cell (250 x 10 mm) through the opening of loop-LR (5). After valve-1 and Loop-LR were switched, the injection valve-2 (7) was opened and a volume of 500 µL of 10% m/v SnCl2 in 6.0 mol L-1 HCl solution was added to reduce the Hg(II) ions to mercury vapor. The mercury vapor generated was carried out with a nitrogen flow by opening the injection valve-3 (8) towards the flow cell of the spectrometer (10), and the signal was recorded in peak height mode.



Results and discussion

Effect of physical and chemical parameters

In order to adjust the physical and chemical variables for mercury preconcentration and elution, the following parameters were tested.

The preconcentration and elution flow rates were varied from 0.5 to 3 mL min-1 while the amount of injected mercury was kept constant at 600 ng L-1. As shown in Fig. 2 (graph a), optimum signal absorbance was obtained at a flow rate of about 1.0 mL min-1. The sensitivity increased with increasing elution flow rate (Fig. 2, graph b.), remaining constant at 1.5 mL min-1. As a compromise between the sampling rate and sensitivity, the preconcentration and elution flow rates selected were 1.0 and 1.5 mL min-1, respectively. Various column lengths and inner diameters were tested, resulting in the selection of 4 cm and 2 mm diameters, respectively, as a compromise between loading capacities and eluant volumes.



The effect of pH on mercury sorption in the column's reactive phase was evaluated by varying the pH in the 1 - 5 range [19] with a constant injection of 600 ng L-1 of mercury solution and 3 min. sampling time. Figure 3 shows that the sorption of Hg(II) ions decreased with the solution's acidity and pH 5.0 was optimal for the sorption of Hg(II) ions. A pH of 5.0 was chosen for further experiments, using 5.10-3 mol L-1 of acetate buffer.



The elution efficiency of Hg(II) ions was investigated using HNO3 and HCl in different concentrations. Fig. 4 depicts the influence of eluant concentrations, indicating that the HCl solutions proved the most favorable for eluting Hg(II) ions. However, the elution efficiency improved as the HCl concentration was increased up to 2 mol L-1. Therefore, the 2 mol L-1 HCl solution was chosen as eluant.



The formation and generation of mercury vapor were investigated by varying the concentrations of reducing agent (5 - 20% m/v SnCl2), the concentrations of hydrochloric acid (1 - 10 mol L-1), the volume of SnCl2 (50 - 1000 µL) and the Hg carrier stream (10 - 50 mL min-1 N2). The results obtained in these experiments, which are illustrated in Fig. 5 and Fig. 6, indicate that, when the N2 flow rate was increased from 10 to 50 mL min-1, the highest absorbance peak for the solution containing 600 ng L-1 Hg(II) was reached with a N2 flow rate of 20 mL min-1 and a 10% m/v SnCl2 volume (in 6 mol L-1 HCl solution) of 500 µL.





Analytical Performance

A linear calibration curves were obtained for the 80 – 800 ng L-1 range of Hg(II) preconcentrated for 3, 4, and 5 min (volumes injected in triplicate at a flow rate of 0.5 mL min-1). The results obtained are illustrated in Figure 7, which indicates that the calibration slopes increased, though not in proportion with the increasing preconcentration time; hence, the retention efficiency was not constant.



The experimental preconcentration factor, calculated as the ratio of slopes in the calibration graphs obtained from preconcentrations in the minicolumn packed with SiAT (A= 9.83 x 10-3 + 1.92 x 10-4[Hg2+]), the preconcentration in the gold minicolumn (A= 4.66 x 10-3 + 9.70 x 10-5[Hg2+] ) and no preconcentration (A= 5.145 x 10-4 + 6.60 x 10-6[Hg2+] ), was evaluated under the same conditions as described previously. Figure 8 shows the aforementioned calibration graphs. The preconcentration factor, calculated by comparing the calibration graph with preconcentrations (Figure 8, graph c) against the calibration graphs without preconcentrations (Figure 8, graph a) and with the preconcentration in the gold minicolumn (Fig.8, graph b), was 29 and 2, respectively. Differences between the preconcentration factor and the added and eluted volumes may occur due to loss of efficiency with increasing sample volumes. As it can be seen in Table 1, the preconcentration efficiency declined from 98% to 92% when the sample volumes passed through the column increased from 3 to 5 mL. The preconcentration factor under these conditions increased from 29 to 46. The precision, expressed as RSD for twelve independent determinations, was 4% for 200 ng L-1 Hg(II) solutions. The detection limit (LOD) calculated on the basis of 3 SD/b, were b is the slope of the calibration curve c illustrated in Figure 8 and SD the standard deviation of 10 consecutive measurements of blank solution, was 5 ng L-1.





The repeatability of column preconcentrations, expressed as SD and tested by sampling a 400 ng L-1 Hg(II) solution for 5 min through 12 independently prepared minicolumns, was approximately 95 ± 6.

The influence of dissolved salts on the preconcentration of Hg(II) was evaluated by adding 500 µg L-1 of Cu(II), Ni(II), Fe(III)/Fe(II), Pb(II), Zn(II), Mn(II), Cd(II), Cl-, NO3-, S042- and PO43-. The interference study (Table 2) showed that the proposed method is fairly selective for Hg(II) and that Cl- caused most of the interference. This was attributed to the formation of chlorocomplexes of mercury (HgCl42-). However, the interference of this anion in water samples can be eliminated by using the digestion procedure proposed in this work. In this procedure the Cl- ions are convert to Cl2 by oxidizing photocatalysis in the presence of 100 mg of TiO2 and 0.01 mol L-1 potassium persulfate. In this conditons the Hg(II) ions to remain stable in the solutions.



Determination of mercury in river water samples

The proposed method was successfully applied in mercury recovery tests using river water samples containing 20 -25 mg L-1 of dissolved organic carbon. The results achieved through the proposed method are illustrated in Table 3, which shows recoveries of up to 97% were attained in predigested samples and 70% in undigested samples. This difference between the recovery results, was attributed to the formation of inert complexes of mercury with aquatic humic substances (HS), which are less accessible for the functional groups of the collector [25]. After digestion (by oxidizing with photocatalysis in the presence of 100 mg of TiO2 and 0.01 mol L-1 potassium persulfate), Hg(II) ions previously bound to HS as inert complexes can react quantitatively with functional groups of SiAT, due the total mineralization of organic matter in the process of digestion by photocatalysis.




The proposed method offers a good alternative for determining mercury in ultratrace levels in an aqueous medium. The high stability of the mini-column packed with SiAT was evidenced by lifetimes of at least one year when unused and at least two month when used 4 h per day (aproximately 800 analysis). The proposed method allows mercury to be determined on-line by CVAAS in natural water samples, but requires predigestion of the sample, e.g., by oxidizing with photocatalysis in the presence of 0.01 mol L-1 potassium persulfate, which is necessary to eliminate the interference of Cl- ions and to inhibit the reduction of Hg(II) by UV radiation during the digestion phase.



The authors thank the Brazilian research funding agencies FAPESP (Project 99/12916-0, 03/13326-6) for the financial support of this work and CAPES for fellowships granted to P. S. Roldan and A. B. Araujo.



[1] E. D., Bidone; Z. C., Castilhos, T. M., Cid., De Zouza, L. D., Lacerda, Bulletin Environmental Contamination and Toxicology 59 (1997) 201.        [ Links ]

[2] E. D., Bidone; Z. C., Castilhos, T. M., Cid., De Zouza, L. D., A screening approach Water, Air and Soil Pollution, 97 (1997) 15.        [ Links ]

[3] V., Hatje, E. D., Bidone, J. L., Maddok, Environmental Technology 17 (1998) 487.        [ Links ]

[4] F. S., Rodrigues, J. E. L., Maddok, Journal Geochem. Explor. Amsterdam 58 (1997) 240.        [ Links ]

[5] D., Amouroux, J. C., Wasserman, E., Tessier, O., Donard, Environmental Science & Technology 33 (1999) 3051.        [ Links ]

[6] C. M., Tseng, A., Diego, J. C., Wasserman, Chemosphere 39 (1999) 1136.        [ Links ]

[7] J. E. Scott, J. M. Ottaway, Analyst 106 (1981) 1081.        [ Links ]

[8] R., Dumarey, E., Temmerman, R., Dams, J., Hoste, Anal. Chim. Acta 170 (1985) 340.        [ Links ]

[9] E. Temmerman, C. Vandecasteele, G. Vermeir, R. Leyman, R. Dams, Anal. Chim. Acta 236 (1990) 376.        [ Links ]

[10] L. Liang, N. S. Bloom, J. Anal. At. Spectrom. 8 (1993) 594.        [ Links ]

[11] M. C. R., Resende, R. C., Campos, A., J., Curtius, J. Anal. At. Spectrom. 8 (1993) 251.        [ Links ]

[12] L. M. Aleixo, M. F. B. Sousa, O. E. S. Godinho, G., Oliveira Neto, Y., Gushikem, J. C., Moreira, Anal. Chim. Acta 271 (1993) 217.        [ Links ]

[13] J. C., Andrade, J. C., M. M. S., Bueno, Spectrochimica Acta, Part B 49B (1994) 795.        [ Links ]

[14] M. C., Canela, W. F., Jardim, J. J. R., Rohwedder, J. Autom. Chem. 18 (1996) 198.        [ Links ]

[15] H. Sakamoto, J. Taniyama, N. Yonehara, Anal. Sci. 13 (1997) 775.        [ Links ]

[16] C. E. C, Magalhães, F. J., Krug, A. H., Forstier, H. Berndt, J. Anal. At. Spectrom. 12 (1997) 1234.        [ Links ]

[17] N. Miekeley, M. O. Amato, Atomic Spectroscopy 18(6) (1997) 191.        [ Links ]

[18] J. A. Gomes Neto, L. F. Zara, A. Santos, J. C. Rocha, A. A. Cardoso, Laboratory Robotics and Automation, 11 (1999) 310.        [ Links ]

[19] J. A. Gomes Neto, L. F. Zara, J. C. Rocha, A. Santos, C. S. Dakuzaku, J. A. Nóbrega, Talanta 51 (2000) 594.        [ Links ]

[20] O. Wurl, O. Elsholz, R. Ebinghaus, Anal. Chim. Acta 438 (2001) 249.        [ Links ]

[21] J. C., Moreira, Y., Gushikem, Anal. Chim. Acta (1985) 267.        [ Links ]

[22] L. T. Kubota, J. c., Moreira, Y. Gushikem, Analyst 114 (1989) 1388.        [ Links ]

[23] N. L. Dias Filho, Y., Gushikem, W. L., Polito, J. C. , Moreira, E. O., Ehrim, Talanta 42 (1995) 1630.        [ Links ]

[24] P. Lessi, N. L. Dias Filho, J. C. Moreira, J. T. S. Campos, Anal. Chim. Acta 327 (1996) 190.        [ Links ]

[25] P. M. Padilha, L. A. M. Gomes, C. C. F. Padilha, J. C. Moreira, N. L. Dias Filho, Anal. Lett. 32 (1999) 1820.        [ Links ]

[26] A. M. Gomes, P. M. Padilha, J. C. Moreira, N. L. Dias Filho, Y. Gushikem, J. Bras. Chem. Soc. 9 (1998) 499.        [ Links ]

[27] A. G. S. Prado, C. Airoldi, Anal. Chim. Acta 432 (2000) 211.        [ Links ]

[28] I. L Alcântara, P. S. Roldan, G. R. Castro, F. V. Moraes, F. A. Silva, C. C. F. Padilha, J. D. Oliveira, P. M. Padilha, Anal. Sci. 20 (2004) 1029-1032.        [ Links ]



Recebido em: 16/03/2005
Aceito em: 19/05/2005



* E-mail:

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License